Marine Design & Research Institute of China

Prevailing quantization techniques in Learned Image Compression (LIC) typically employ a static, uniform bit-width across all layers, failing to adapt to the highly diverse data distributions and sensitivity characteristics inherent in LIC models. This leads to a suboptimal trade-off between performance and efficiency. In this paper, we introduce DynaQuant, a novel framework for dynamic mixed-precision quantization that operates on two complementary levels. First, we propose content-aware quantization, where learnable scaling and offset parameters dynamically adapt to the statistical variations of latent features. This fine-grained adaptation is trained end-to-end using a novel Distance-aware Gradient Modulator (DGM), which provides a more informative learning signal than the standard Straight-Through Estimator. Second, we introduce a data-driven, dynamic bit-width selector that learns to assign an optimal bit precision to each layer, dynamically reconfiguring the network's precision profile based on the input data. Our fully dynamic approach offers substantial flexibility in balancing rate-distortion (R-D) performance and computational cost. Experiments demonstrate that DynaQuant achieves R-D performance comparable to full-precision models while significantly reducing computational and storage requirements, thereby enabling the practical deployment of advanced LIC on diverse hardware platforms.

With the aggravation of global environmental problems and the continuous promotion of the "double carbon" goal, new energy and electrified transportation have developed rapidly. Electric vehicles (EVs), as a novel type of mobile load or power source, exhibit more pronounced randomness and clustering characteristics compared to photovoltaics, wind turbines, and energy storage systems. The large-scale integration of EV into the power grid significantly alters fault characteristics on protection lines, adversely affecting the performance of current differential protection. This study first explores the basic topology of EV charging stations and then, based on the principle of ratio-restrained current differential protection, proposes guidance strategies to ensure the reliability and safety of grid operations under typical EV discharge scenarios. By reasonably controlling the discharging power of EV, the sensitivity and reliability of current differential protection are effectively enhanced. Finally, a large-scale EV discharging simulation model is developed on the PSCAD/EMTDC platform using a modified IEEE 33-bus test system, validating the impact of EV discharging on differential protection and the effectiveness of the proposed guidance strategies.

With the increase in carbon emissions and climate change, natural disasters are occurring more and more frequently. As the end of the power grid that delivers electric power to consumers, if the distribution network is affected by disasters, it will lead to load power outages, seriously affecting the safety and reliability of the power supply. To address this issue, an energy storage optimal allocation method is proposed to enhance the resilience of the distribution network. Firstly, a three-stage optimization model of Defense-Attack-Defense (DAD) is established. Through the optimal allocation of energy storage before disasters and the network reconfiguration (NR) method after disasters occur, the load loss under multiple disaster scenarios is minimized. For this multi-scenario model, this paper uses the progressive hedging algorithm (PHA) to solve it. A case study is carried out on the modified IEEE-33 bus system. The results show that the proposed method can effectively improve the resilience of the distribution network under disasters and reduce the risk of load loss.